Unverified50% confidenceFactExact time
The core idea of LLM-as-a-judge is to use a more capable large language model (typically GPT-4, Claude, etc.) to score or rank the outputs of another model.
1
Sources
50%
Confidence
Long-term
Relevance
9/20/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedLLM as Judge was systematically proposed by Lianmin Zheng et al. in the MT-Bench paper (2023), with the core idea of using an independent large language model to score the output of another model as a substitute for human annotation.73% similarVerifiedLarge language models (LLMs) such as GPT and Claude are essentially probabilistic prediction systems trained on massive amounts of text data.73% similarVerifiedLLM-as-a-judge模式即用一个能力更强的模型来评判另一个模型的输出质量，降低评测成本但引入评判模型自身的偏见问题73% similarUnverifiedLLM evaluation requires handling semantic similarity rather than exact matching, so evaluation frameworks themselves also rely on language models as judges (LLM-as-a-Judge).71% similarUnverifiedScoring functions often require careful design and may even rely on another LLM for scoring, known as the LLM-as-judge pattern.70% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/936115API
curl https://kongchang.com/api/v1/knowledge/claims/936115MCP
get_claim(id=936115)